Abhijit Tripathy started Presear Softwares from a Central University incubator with no office, no co-founder, and no paying client. Six years later, he has 150+ AI deployments, 100+ enterprise clients across 20+ sectors, and a bold claim: that India’s most sensitive institutions — defence, government, healthcare — should run on infrastructure built in India, not rented from America.
Starting From Scratch in Bilaspur
Abhijit Tripathy didn’t launch Presear Softwares with venture backing, a co-founder, or even a desk of his own. He started in 2019 out of the GGV Incubation Centre at Guru Ghasidas Vishwavidyalaya in Bilaspur, Chhattisgarh — armed with a technical framework he’d named FUMER (Functional, Usable, Maintainable, Efficient, Reliable) and a frustration he couldn’t shake.
The frustration was simple, but the problem ran deep. Indian enterprises were caught in a trap: either adopt global SaaS tools built with no understanding of Indian languages or regulatory realities, or commission expensive custom builds that dragged on for years and rarely survived past the first demo. Neither path produced AI that genuinely transformed how an organisation worked. Tripathy believed there was a third option — and he set out to build it.
Doing the Unglamorous Work First
Presear’s early projects weren’t the stuff of startup mythology. Government contracts. Smart city systems. Processing tens of thousands of applications for urban local bodies. Unremarkable on paper, but exactly the kind of work that builds real institutional credibility — and more importantly, reveals what enterprise clients actually need versus what they say they need. Those patterns, accumulated over years of delivery, eventually became the foundation for Presear’s product suite.
A Platform, Not a Point Solution
Today, Presear offers five products designed to work together as a unified stack. DeepQuery Engine makes internal document repositories truly searchable and queryable in real time. IndiCompute Cloud delivers sovereign, on-premise AI infrastructure for organisations where data leaving the country isn’t a theoretical concern — it’s a compliance violation. PresearBoloAI and its companion product Dakshini bring conversational AI to over 55 Indian languages, including regional variants, with domain-specific training tailored to banking, healthcare, and legal use cases. GentrikOS handles AI-driven industrial and operational workflows. And PresearHELIX serves as the governance layer across everything — policy enforcement, audit trails, identity management.
What ties them together is an architectural philosophy: the same underlying stack, configured differently for different sectors, rather than a separate build for each client. That’s what lets a company of Presear’s size serve defence, healthcare, fintech, legal, and government clients simultaneously without spreading itself thin.
The most pointed product in the portfolio is IndiCompute Cloud. Tripathy is diplomatic about the hyperscalers — he acknowledges that AWS and Azure are technically excellent — but his argument isn’t about capability. It’s about structure. For certain sectors and certain categories of data, the question of where infrastructure physically lives and under whose legal jurisdiction it operates isn’t a preference. It’s a requirement. IndiCompute is built specifically for that reality: private and on-premise deployment, sub-100ms inference performance, and full data residency compliance. The goal isn’t to compete with Amazon globally. It’s to be the obvious answer for Indian defence, government, and regulated industries.
Competing on Accountability
When people ask how a Bhubaneswar-based startup wins enterprise contracts against TCS or Infosys, Tripathy’s answer tends to catch them off guard. He doesn’t lead with price. He leads with accountability.
Large IT firms bring considerable resources to an enterprise engagement — and then distribute responsibility across a rotating roster of consultants, many of whom weren’t present when decisions were made six months earlier. Presear’s pitch is different: production-grade AI delivered in weeks rather than 18 months, with the same people who built the system available when something needs fixing. That kind of direct accountability is genuinely hard for large organisations to replicate. Their scale is both their selling point and their structural weakness.
Built on Grants, Not Dilution
Presear has raised over $386,000 through three Government of India schemes — DST TEC, DST NIDHI i-TBI, and MSME TBI — and has received recognition from Startup Odisha. For Tripathy, this wasn’t just opportunistic. It was a deliberate choice to avoid giving away equity during the company’s most vulnerable early years, when valuation is weakest and leverage is lowest. Grant funding bought time — to build product, develop intellectual property, publish research, and accumulate the 150+ deployments and 100+ enterprise clients that now make the revenue story credible to institutional investors.
The current model is consulting-led, which funds ongoing R&D and doubles as a discovery engine for productisable patterns. Over the next three years, Tripathy expects SaaS revenue from DeepQuery, BoloAI, and GentrikOS to become the dominant line, with consulting evolving into a premium deployment layer rather than the core offering.
The Odisha Bet
Presear now has offices in Bhubaneswar, Bengaluru, and Pune. But the headquarters, the founding story, and the cultural identity are rooted in Odisha — intentionally so. Tripathy wants to make a practical argument, through the company’s actual existence, that serious AI infrastructure can be built outside India’s Tier-1 tech corridors. The advantages he points to are concrete: proximity to state government relationships, significantly lower operating costs than Bengaluru’s prime tech districts, and a team that chose to be there because of the mission rather than the postcode.
On governance and ethics, Presear holds a Carbon Neutral certification and built PresearHELIX to operationalise policy enforcement and audit infrastructure from day one. Tripathy’s view is that in India’s current regulatory environment, having this infrastructure already built is a competitive advantage — particularly for global enterprise clients and MNC subsidiaries operating in India, who are watching the regulatory direction closely and don’t want to retrofit governance later.
What He’s Actually Building Toward
Ask Tripathy about the ten-year picture and he’s unusually direct about the fact that an exit isn’t the primary frame. What he’s optimising for is a condition: one in which Indian AI infrastructure — for enterprises, for government institutions, for defence — is actually built in India, operated by Indian organisations, and owned by an Indian company. Not licensed from an American platform. Not assembled from global components with an Indian label on top. The infrastructure itself.
Whether that path ends in a public listing or something else is, in his own words, secondary. The goal is the thing itself. From a room in a Central University incubator in Bilaspur, that once sounded like an outsized ambition. Six years and 150 deployments later, it sounds more like a roadmap.
Interview By: Arushi Agarwal
